VM Resource Optimization via Machine Learning Placement

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current systems lack an efficient method to optimize the configuration of virtual machines (VMs) and containers on nodes with varying hardware resources, leading to suboptimal resource utilization and inefficient deployment in virtual cloud platforms.

Innovation Solution

The use of machine learning techniques to determine optimal VM configurations based on input parameters, employing methodologies like 'stacking' and 'spreading' to place containers on nodes, and generating configuration scores for resource utilization and performance metrics, allowing for the selection and deployment of optimal configurations using an orchestration system.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If traditional resource allocation methods are used to deploy containers on nodes, then deployment simplicity is maintained, but resource utilization efficiency deteriorates

Engineering Contradiction:

Improveresource utilization efficiencyVSAvoiddeployment complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The system employs machine learning models that automatically analyze container requirements and node characteristics to self-determine optimal deployment configurations without manual intervention. The orchestration system autonomously evaluates multiple factors including hardware resources, software dependencies, and performance metrics to make intelligent placement decisions, thereby improving resource utilization while maintaining operational simplicity.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces traditional manual or rule-based resource allocation mechanisms with machine learning-based intelligent systems. The ML models process complex multi-dimensional data about containers and nodes to generate optimized deployment configurations, substituting mechanical deployment processes with data-driven automated decision-making that achieves superior resource utilization efficiency.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Adaptability or versatility

If more nodes are provisioned to accommodate varying container resource requirements, then container deployment flexibility is improved, but hardware resource waste increases

Engineering Contradiction:

Improvecontainer deployment flexibilityVSAvoidhardware resource waste

Core Design Contradiction:

Adaptability or versatilityVSLoss of energy

Solution Approach 1:

The system dynamically adjusts deployment parameters based on real-time analysis of container resource requirements and node availability. Machine learning models continuously evaluate changing conditions and optimize configuration parameters such as resource allocation ratios, placement strategies, and scaling factors to achieve optimal resource utilization while maintaining deployment flexibility for varying container workloads.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The orchestration system provisions nodes and resources on-demand based on actual container deployment needs rather than pre-configuring fixed infrastructure. By implementing partial provisioning strategies where resources are allocated only when and where needed, the system maintains high adaptability for diverse container requirements while minimizing hardware resource waste through selective and efficient resource utilization.

Inventive Principle:

Principle #16Partial or excessive action

3Productivity

If manual configuration optimization is performed to improve resource allocation, then resource utilization improves, but time consumption and operational complexity increase

Engineering Contradiction:

Improveresource allocation efficiencyVSAvoidconfiguration optimization time

Core Design Contradiction:

ProductivityVSLoss of time

Solution Approach 1:

The system performs preliminary analysis and pre-computation of optimal deployment configurations using machine learning models trained on historical deployment data. By pre-processing container requirements and node characteristics to generate optimized placement recommendations in advance, the system achieves efficient resource allocation automatically without requiring time-consuming manual optimization efforts during actual deployment operations.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The orchestration system implements continuous feedback loops that monitor actual resource utilization and deployment performance, using this information to refine and update machine learning models. The feedback mechanism automatically adjusts configuration strategies based on observed outcomes, enabling the system to improve resource allocation efficiency over time without manual intervention while reducing the need for repetitive configuration optimization tasks.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS11579939B2Systems and methods for virtual machine resource optimization using machine learning techniques

Publication Date: 2023.02.14 VERIZON PATENT & LICENSING INC

AI summary

Systems described herein may allow for the intelligent configuration of containers onto virtualized resources. As described, systems described herein may generate configurations based on received parameters for utilization to configure (e.g., install, instantiate, etc.) virtualized resources. Once generated, a configuration may be selected according to determined selection parameters and/or intelligent selection techniques.